Multiple genome alignment in the telomere-to-telomere assembly era

Abstract

With the arrival of telomere-to-telomere (T2T) assemblies of the human genome comes the computational challenge of efficiently and accurately constructing multiple genome alignments at an unprecedented scale. By identifying nucleotides across genomes which share a common ancestor, multiple genome alignments commonly serve as the bedrock for comparative genomics studies. In this review, we provide an overview of the algorithmic template that most multiple genome alignment methods follow. We also discuss prospective areas of improvement of multiple genome alignment for keeping up with continuously arriving high-quality T2T assembled genomes and for unlocking clinically-relevant insights.

Keywords: Comparative genomics; Homology; Multiple genome alignment; Synteny.

Fig. 1

An example of evolution resulting in different classes of homology. In (a), deletion and speciation events are depicted, resulting in both orthologs and paralogs. The resulting relationships between all pairs of X segments as well as between all pairs of Y segments are depicted in (b). Notably, segments $Y_b^3$ and $Y_c^2$ participate in three important homology relationship types. As the most recent common ancestor of $Y_a^3$ and $Y_b^3$ is at a speciation event, they are orthologous to each other. Segments $Y_c^2$ and $Y_a^1$ on the other hand are paralogs, as their homology is a result of duplication. Finally, $Y_b^3$ and $Y_c^2$ are special types of paralogs known as “false orthologs” i.e. paralogous segments all of whose other copies in their respective genomes have since been deleted, resulting in two segments which appear to never have been duplicated. The presence of such false orthologies complicates the problem of core-genome alignment particularly, but also the problem of further categorizing homologies identified through genome alignment. It is worth noting that homology relationships within a single species’ genome do exist, but are not depicted here. For example, $Y_a^1,Y_a^2$ and $Y_a^3$ are all paralogous to each other

Fig. 2

Timeline of MGA tools. The original sequencing of human and mouse genomes spurred the development of a number of multiple genome alignment tools. Following this initial spurt, the next generation of genome aligners (starting with Enredo-Pecan and ending with Cactus) were developed, followed by a 6-year period of silence, with Parsnp being one of the few tools released between 2012 and 2019

Fig. 3

a-c The number of assemblies available for different groups of eukaryotic species in NCBI. d-f The number of eukaryotic species with available genome assemblies in NCBI. The emergence of 3rd-generation sequencing technologies and their accompanying assembly algorithms in the early 2010s is largely responsible for the increasing rate of novel eukaryotic genome deposits

Fig. 4

The pipeline of MGA. a An example of large-scale genomic rearrangement, insertion, and deletion occurring across a set genomes. The lines denote bounds of homologous segments, and inversions are denoted by crossing lines. b Anchors from the section of 3 genomes surrounded by the dotted box in a. Once again, the lines between genomes represent homology. The labeled blocks on each genome correspond to anchors, where two blocks with the same label are inferred to be potentially homologous sites. c The alignment graph obtained by merging the 3 linear genome graphs in b. At this step, the aim of MGA is to find colinear paths in the graph i.e. sequences of anchors which are traversed by a group of genomes in the same order. d Often times, the initial set of anchors will be too noisy, containing spurious alignments which prevent the formation of longer, more reliable colinear paths. By removing anchor F, the alignment graph becomes much simpler and yields longer colinear paths, where each set of colinear paths is denoted by the color. e For each colinear path, MGA tools perform an MSA, yielding a set of sequence alignments which together make up the genome alignment